No surprise here. This kind of query is known to be slow. Moreover, the way you constructed the query would even return incorrect results, as you wouldn't get all communities in your department (only those up to 10km proximity to the border).
So we use a completely different approach and split **Note: I had to remove most parts of my previous answer, as it didn't properly take the query in two halves:
10km distance into account. Sorry for that.**
1. Find all Up to now, the query would simply return any adjacent boundary=administrative with admin_level = 8 inside 8.
I've submitted a [pull request](https://github.com/drolbr/Overpass-API/pull/167) to improve the department (.rel1)
2. Find all nodes up to 10km distance to the border of the department, and for those nodes, find the corresponding community they are around performance. With that fix in (boundary=administrative and admin_level = 8) (.rel2).
We merge the result and display it using new geometry (out geom).
This query takes just about 30s (whereas place, your query ran for at least for 10 minutes without any result). It returns will run in about 20MB of data, so careful with your browser.
For the impatient, here's a screenshot of the department border along with the communities (inside the department and up to 10km distance to the border).
![Note: Combined image with department + communities borders was made with gimp, based on two different Overpass queries.][2]
Overpass QL:
area[name="Ille-et-Vilaine"][boundary=administrative];
rel(area)[boundary=administrative]['admin_level'='8']->.rel1;
rel[name="Ille-et-Vilaine"][boundary=administrative];
>;
node._(around:10000);
is_in;
area._[boundary=administrative]['admin_level'='8'];
rel(pivot)->.rel2;
(.rel1; .rel2;);
out geom;
2 minutes.
Link to previous (non-working) overpass turbo: turbo query: http://overpass-turbo.eu/s/6oX
Remember: *Don't forget to come over and cast your vote.*
[1]: http://help.openstreetmap.org/upfiles/op2.png
[2]: http://help.openstreetmap.org/upfiles/op2_6H0u9dc.png
